Method

MonoCInIS: Camera Independent Monocular 3D Object Detection using Instance Segmentation [MonoCInIS]


Submitted on 6 Oct. 2021 19:06 by
Jonas Heylen (KU Leuven)

Running time:0,14 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
In this paper we propose a category-level pose
estimation method based on instance segmentation,
using camera independent geometric reasoning to
cope
with the varying camera viewpoints and intrinsics
of
different datasets. Every pixel of an instance
predicts the object dimensions, the 3D object
reference points projected in 2D image space and,
optionally, the local viewing angle. Camera
intrinsics are only used outside of the learned
network to lift the predicted 2D reference points
to
3D.
Parameters:
2RP method
Latex Bibtex:
@inproceedings{heylen2021monocinis,
title={MonoCInIS: Camera Independent Monocular
3D Object Detection using Instance Segmentation},
author={Heylen, Jonas and De Wolf, Mark and
Dawagne, Bruno and Proesmans, Marc and Van Gool,
Luc and Abbeloos, Wim and Abdelkawy, Hazem and
Reino, Daniel Olmeda},
booktitle={Proceedings of the IEEE/CVF
International Conference on Computer Vision},
pages={923--934},
year={2021}
}

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Car (Detection) 96.22 % 88.16 % 75.72 %
Car (Orientation) 45.00 % 40.75 % 34.48 %
Car (3D Detection) 15.21 % 7.66 % 6.24 %
Car (Bird's Eye View) 20.42 % 10.96 % 9.23 %
This table as LaTeX


2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot



3D object detection results.
This figure as: png eps pdf txt gnuplot



Bird's eye view results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker